[oor-forum] oor_20140722: Chat Transcript - sent by: TillMossakowski

[Till Mossakowski]

*Chat transcript from room: oor_20140722*
*2014-07-22 GMT-08:00*
*[07:03] **TillMossakowski: *wiki page:
http://ontolog-02.cim3.net/wiki/OOR_Conference_2014_07_22 Skype:
join.conference when prompted enter Conference ID: 843758#
*[07:17] **TillMossakowski: *slides also available at
http://iws.cs.uni-magdeburg.de/~mossakow/2014-07-22_Ontohub_API.pdf
*[07:19] **ToddSchneider: *Does the generated specification depend on
the coding language?
*[07:28] **ToddSchneider: *So a query is not treated as an object?
*[07:30] **ToddSchneider: *The description of ontologysearch is not
clear to me.
*[07:31] **TillMossakowski: *slide 13...
*[07:34] **ToddSchneider: *What is the context of the APIs? Instances of
ontohub (i.e., inter-ontohub)? End users? Both?
*[07:36] **ToddSchneider: *Are 'accepts', 'registers with', 'informs'
API commands?
*[07:37] **ToddSchneider: *How does 'reliance' handle end-point or
communication failures?
*[07:41] **TillMossakowski: *Todd: does the API cover authentification?
Tim: yes, using an authorization token approach.
*[07:42] **MikeBennett: *This might be a naive question: can these API
commands be executed from the command line? So that for instance a
scheduling tool like Jenkins could invoke them?
*[07:45] **TillMossakowski: *Tim's answer: yes.
*[07:46] **MikeBennett: *Can you search on ontology imports i.e. get a
result showing how ontologies import from one another?
*[07:46] **TillMossakowski: *Todd: Does the generated specification
depend on the coding language? Tim: we use ruby, and the DSL will be
based on ruby.
*[07:51] **TillMossakowski: *Tim: in the API, ontology-related queries
use the IRI of the ontology, followed by a command. Other queries
consist of just the command. In both cases, there may be additional
parameters.
*[07:51] **TillMossakowski: *Todd: why constrain this prefixing approach
just to ontologies?
*[07:52] **TillMossakowski: *Todd: the general approach to use a IRI as
a prefix that gives the context to the subsequent command is interesting.
*[07:53] **TillMossakowski: *Todd: the prefix approach could be used for
any resource (text files, image files, ...), not just ontologies.
*[07:54] **ToddSchneider: *Can imports be considered as a 'type' of
[ontology] mapping?
*[07:55] **TillMossakowski: *Mike: Can you search on ontology imports
i.e. get a result showing how ontologies import from one another? Tim:
this would be a textual representation of the ontology graph. Would be
interesting to add this.
*[07:57] **ToddSchneider: *Would it make sense to include a 'resource
type' as part of the prefix?
*[07:57] **TillMossakowski: *Tim: until now, we compute all query
answers on the fly, later on, caching will be needed.
*[07:57] **TillMossakowski: *Tim: imports are a kind of ontology mapping.
*[07:58] **RaviSharma: *thanks for answering search during import on the
fly.
*[08:00] **MikeBennett: *I have to drop off now
*[08:01] **TillMossakowski: *Todd: Would it make sense to include a
'resource type' as part of the prefix? Tim: yes, we should think about
this, to prefix or postfix the IRI with the type, in order to ensure
that the command is relevant in the given context.
*[08:04] **ToddSchneider: *Thank you.
*[08:04] **TillMossakowski: *Agreement: in a few weeks, have will a
follow-up meeting, with discussion of more details.
*[08:04] **RaviSharma: *thanks
*[08:05] **TillMossakowski: *Next week: general OOR strategy discussion.
*[08:05] **TimReddehase: *Thank you all for the suggestions and
revealing questions.
*[08:05] *List of attendees: Gary Berg-Cross, MichaelGruninger,
MikeBennett, MikeDean, RaviSharma, TillMossakowski, Tim, TimReddehase,
ToddSchneider
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.ovgu.de/pipermail/oor-forum/attachments/20140723/386197dc/attachment.html>